Interference with Phosphoenzyme Isomerization and Inhibition of the Sarco-endoplasmic Reticulum Ca 2 (cid:1) ATPase by 1,3-Dibromo-2,4,6-tris(methylisothiouronium) Benzene*

ATP hydrolysis and Ca 2 (cid:1) transport by the sarco-endo-plasmic reticulum Ca 2 (cid:1) ATPase (SERCA) are inhibited by 1,3-dibromo-2,4,6-tris(methylisothiouronium) benzene (Br 2 -TITU) in the micromolar range (Berman, M. C., and Karlish, S. J. (2003) Biochemistry 42, 3556–3566). In a study of the mechanism of inhibition, we found that Br 2 -TITU allows the enzyme to bind Ca 2 (cid:1) and undergo phosphorylation by ATP. The level of ADP-sensitive phosphoenzyme ( i.e. E 1P-2Ca 2 (cid:1) ) observed in the tran-sient state following addition of ATP is much higher in the presence than in the absence of the inhibitor. Br 2 - TITU does not interfere with enzyme phosphorylation by P i in the reverse direction of the cycle ( i.e. E 2P) and produces only a slight inhibition of its hydrolytic cleavage. The inhibitory effect of Br 2 -TITU on steady state ATPase velocity is attributed to interference with the E 1P-2Ca 2 (cid:1) to E 2P-2Ca 2 (cid:1) transition. In fact, experiments on conformation-dependent protection from proteolytic digestion suggest that, in the presence of Br 2 -TITU, the loops connecting the “A”

The sarco-endoplasmic reticulum Ca 2ϩ ATPase (SERCA) 1 is a membrane-bound 100-kDa protein (2) that sustains active transport of Ca 2ϩ , coupled to utilization of ATP. The catalytic and transport cycle includes a number of sequential steps (3) beginning with activation of the enzyme ground state (E2) by high affinity binding of 2 Ca 2ϩ (E1-2Ca 2ϩ ) on one side of the membrane followed by utilization of ATP to form of a phospho-rylated enzyme intermediate (ADP-E1P-2Ca 2ϩ ). The free energy derived from ATP is then utilized for isomerization of E1P-2Ca 2ϩ to E2P-2Ca 2ϩ , whereby the bound Ca 2ϩ dissociates with lower affinity on the opposite side of the membrane. Finally, the cycle is completed by hydrolytic cleavage of P i from E2P (Scheme I).
Changes of protein conformation, including separation, rotation, and gathering of the three (N, P, and A) cytosolic domains as well as the displacement of transmembrane segments, occur in concomitance with the sequential steps of the ATPase cycle and play an essential role in the mechanism of energy transduction (4 -8). In fact, SERCA inhibition can be produced by stabilization of specific conformational states. For example, thapsigargin (TG) binds to the ATPase with high affinity, yielding a dead end complex with the enzyme in ground state (9,10). We report here a series of experiments on the characterization of the inhibitory mechanism of Br 2 -TITU (1) and its comparative features relative to other SERCA inhibitors such as TG and 2,5-di(tert-butyl)hydroquinone (DBHQ).
Ca 2ϩ binding in the absence of ATP was measured by incubating SR vesicles (40 g/ml) in a reaction mixture containing 20 mM MOPS, pH 7.0, 80 mM KCl, 5 mM MgCl 2 , 0.2 mM EGTA, and [ 45 Ca]CaCl 2 to yield various concentrations of free Ca 2ϩ and 5 M A23187 ionophore. TG (1 M) was added to half of the samples to provide controls exhibiting no specific Ca 2ϩ binding because it was demonstrated previously that TG prevents specific Ca 2ϩ binding (9,10). The reaction temperature was 25°C.
Enzyme phosphorylation by ATP was measured in an ice-cold reaction mixture containing 50 mM MOPS, pH 7, 80 mM KCl, 2 mM MgCl 2 , 50 M CaCl 2 , 1 mg of SR protein/ml, and 5 M A23187 ionophore. Individual samples (0.2 ml) were started by the addition of 10 M [␥-32 P]ATP and quenched at various times with 1 M perchloric acid.
Enzyme phosphorylation with P i was measured at 25°C in a reaction mixture containing 50 mM MES, pH 6.2, 10 mM MgCl 2 , 2 mM EGTA, various concentrations of [ 32 P]P i and 1 mg of SR protein/ml in a total volume of 0.2 ml. The reaction was acid-quenched after 10 min of incubation by the addition of 1 M perchloric acid.
For all phosphoenzyme measurements, the quenched reaction mixture was transferred into ice, and 1 mg of bovine serum albumin/mg microsomal protein was added as carrier. The samples were then washed by repeated centrifugations and resuspensions in 0.125 M perchloric acid and finally dissolved in 1% SDS for determination of radioactivity and residual protein.
* This work was supported by the NHLBI, National Institutes of Health Grant RO1 HL69830. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Limited proteolytic digestion was performed in reaction mixtures containing 50 mM MOPS, pH 7.0, 50 mM NaCl, 0.6 mg of microsomal protein/ml and 0.02-0.04 mg of proteinase K/ml or 0.02 mg of trypsin/ ml. CaCl 2 , EGTA, TG, or Br 2 -TITU were added as indicated in the figures. Following incubation at 25°C for various time intervals up to 60 min, the reaction was quenched with trichloroacetic acid (2.5%), and the protein was solubilized with SDS (1%), MOPS (0.312 M), pH 6.8, sucrose (3.75%), ␤-mercaptoethanol (1.25 mM), and bromphenol blue (0.025%). The samples were then subjected to electrophoretic analysis (14) on 12% gels followed by staining with Coomassie Blue. In some cases, Western blots were obtained with the monoclonal antibody MA3-911 (Affinity Bioreagents) followed by goat anti-mouse IgG horseradish peroxidase-conjugated secondary antibodies and visualization with an enhanced chemiluminescence-linked detection system (Amersham Biosciences).

RESULTS
In agreement with Berman and Karlish (1), we found that steady state Ca 2ϩ ATPase activity is inhibited by Br 2 -TITU within the micromolar concentration range, yielding a K I of 20 M (Fig. 2). As opposed to the total inhibition produced by TG (9, 10), inhibition by Br 2 -TITU is not total but reduces the steady state ATPase activity by ϳ80%.
Although the previous study of the mechanism of Br 2 -TITU was done by following the development of TNP-AMP superfluorescence upon enzyme activation (1), we directly characterized the ATPase partial reactions by measurements of radioactive isotopes. To this aim, we first measured Ca 2ϩ binding (reaction 1 in Scheme I) in the absence of ATP. We found that Ca 2ϩ binding is not inhibited by Br 2 -TITU, whereas total inhibition is obtained with TG and DBHQ (Fig. 3).
When ATP is added to the enzyme activated by Ca 2ϩ , a phosphoenzyme intermediate is formed rapidly and reaches a level, which is dependent on the rates of phosphoryl transfer from ATP and subsequent hydrolytic cleavage. The experiments shown in Fig. 4 were performed with leaky vesicles (because of the addition of a Ca 2ϩ ionophore) to prevent Ca 2ϩ accumulation in the lumen of the vesicles and consequent inhibition of phosphoenzyme cleavage. Under these conditions, we observed an early rise of the phosphoenzyme followed by reduction to a lower level as the ATP added (10 M) to the reaction mixture was exhausted. Interestingly, we noted that in the presence of Br 2 -TITU the early rise of the phosphoenzyme reached a much higher level and decayed at a slower rate (Fig. 4A). On the other hand, if we added excess ADP 1 s after ATP (i.e. to the peak level of the phosphoenzyme), we observed a rapid decay of the phosphoenzyme both in the absence and in the presence of Br 2 -TITU (Fig. 4B). This indicates that the phosphoenzyme accumulated in the presence of Br 2 -TITU is ADP-sensitive and, therefore, still in the E1P-2Ca 2ϩ state. It is noteworthy that no phosphoenzyme formation with ATP was obtained in the presence of either TG or DBHQ (Fig. 4A).
We then tested the effect of Br 2 -TITU on enzyme phosphorylation by P i in the absence of Ca 2ϩ , which yields equilibrium levels of phosphoenzyme through reactions 7 and 6 in the reverse direction of the cycle (Scheme I). We found no inhibition by Br 2 -TITU, whereas the reaction was strongly inhibited by TG and DBHQ (Fig. 5A). In fact, the phosphoenzyme levels were increased by Br 2 -TITU when low concentrations of P i were used in the reaction mixture, indicating increased affinity of the enzyme for P i in the presence of Br 2 -TITU. When we added excess non-radioactive phosphate to equilibrium levels of the phosphoenzyme obtained with radioactive phosphate, we found that the decay (i.e. hydrolytic cleavage) of the radioactive phosphoenzyme was only slightly inhibited by Br 2 -TITU (Fig. 5B).
It is noteworthy that in the presence of Br 2 -TITU, Ca 2ϩ binding and enzyme phosphorylation with P i were still inhibited by TG or DBH (not shown). Therefore, Br 2 -TITU does not interfere with the binding of these inhibitors and their effects. Furthermore, we found that Br 2 -TITU had the unexpected effect of partially reversing the Ca 2ϩ inhibition of enzyme phosphorylation with P i . Furthermore, Ca 2ϩ binding was reduced in proportion to the enhancement of enzyme phosphorylation (Table I), indicating that the phosphoenzyme formed is in the low Ca 2ϩ affinity state (i.e. E2P).
The kinetic studies described above indicate that the ATPase steady state velocity is inhibited by Br 2 -TITU by interference with an isomeric transition of the phosphorylated enzyme intermediate, whereby conversion of E1P-2Ca 2ϩ to E2P-2Ca 2ϩ is delayed. We then performed experiments on limited ATPase protein digestion with proteinase K and trypsin to explore ligand-dependent conformational states of the enzyme. In these experiments we tested the accessibility of proteinase K (Leu-119 and Thr-242) and trypsin (Arg-198) digestion sites. These sites reside on the loops connecting the "A" domain to transmembrane segments M2 and M3 and on the outer loop of the A domain, respectively. The accessibility of these sites, or lack thereof, reflects A domain rotation in concomitance with sequential reactions of the catalytic cycle and corresponds to dissociation or gathering of the ATPase headpiece domains (15,16).
It is shown in Fig. 6 that, in the absence of Ca 2ϩ , digestion by proteinase K occurs with a specific pattern (17) yielding a 95-kDa band (cleavage at Leu-119) and an 83-kDa band (cleavage at Thr-242). This pattern of rather slow digestion is attributed to the compact conformation of the ATPase headpiece in the absence of Ca 2ϩ (E2) and the consequent protection of A domain proteolytic sites. When the digestion is performed in the presence of Br 2 -TITU, the 95-kDa band is hardly noted, indicating that the Thr-242 site in the A domain is not protected, and the 95-kDa fragment is rapidly cut to yield the 83-kDa product. The opposite is observed in the presence of TG, as a prominent 95-kDa band indicates very efficient protection of the Thr-242 site.
The ATPase digestion with proteinase K proceeds faster in the presence of Ca 2ϩ , and the 95-kDa band is not present either in the absence or in the presence of Br 2 -TITU (Fig. 6). This is attributed to a Ca 2ϩ -dependent, open conformation of the ATPase headpiece (i.e. E1-2Ca 2ϩ ) whereby the proteolytic sites of the A domain are exposed (4). It is noteworthy that under   these conditions the digestion is slower in the presence of TG, with a prominent 95-kDa band (Fig. 6). These experiments indicate that Br 2 -TITU has no apparent influence on the open conformation of the ATPase headpiece, which is induced by Ca 2ϩ (i.e. E1-2Ca 2ϩ ). On the contrary, TG produces stabilization of the E2 conformation and places a strong limit on the fluctuations of the A domain loops. Protection from proteinase K is also observed in the presence of P i and absence of Ca 2ϩ (i.e. E2P). In this case the protection is even higher than in E2, as the ATPase band is reduced at a lower rate, and the two digestion bands are retained for a longer time (Fig. 7). A distinctive feature of the pattern obtained in the presence of Br 2 -TITU is the lack or low intensity of the 98-kDa band (Fig. 7), indicating that the Thr-242 site is not protected. Considering that Br 2 -TITU allows (and even favors, at low P i concentrations) the phosphorylation reaction, it is apparent that considerable fluctuations of the A domain loops are still permitted when the enzyme is phosphorylated by P i in the presence of Br 2 -TITU. On the other hand, the digestion pattern observed in the presence of TG (Fig. 7) is very similar to that observed in the absence of P i , consistent with interference with the P i reaction and stabilization of the E2 state.
Further experiments on proteolytic digestion were performed with trypsin. Trypsin cleaves the ATPase at a first site (T1, Arg-505 on the N domain), yielding two fragments of nearly equal size (A and B), and then at a second site (T2, Arg-198 on the outer loop of the A domain), yielding two subfragments of quite different size (A1 and A2). Digestion at T1 is not influenced significantly by the presence of Ca 2ϩ or P i , whereas digestion at T2 occurs at a rapid rate in the presence of Ca 2ϩ , as revealed by appearance of the A1 band in electrophoretic gels (Fig. 8). Digestion at T2 occurs much more slowly in the absence of Ca 2ϩ and even more slowly under conditions of enzyme phosphorylation by P i (15,16). We found that Br 2 -TITU does not change the patterns of digestion observed in the absence or presence of Ca 2ϩ or P i (Fig. 8). Therefore, Br 2 -TITU does not interfere with conformation-dependent protection of T2 on the outer loop of the A domain. study of fluorescence effects. They observed that Br 2 -TITU quenches the intrinsic tryptophan fluorescence of the ATPase protein but increases the fluorescence acquired by bound TNP-AMP upon enzyme phosphorylation. These effects were attributed to inhibition of phosphoenzyme hydrolytic cleavage and increased levels of the phosphoenzyme in the E2P state (1).
We confirmed that Br 2 -TITU inhibits steady state ATPase activity within the micromolar concentration range with a K I of 20 M. To characterize the mechanism of inhibition, we then directly measured Ca 2ϩ binding in the absence of ATP, and found that Br 2 -TITU had no effect on either the affinity or the stoichiometry of binding. Furthermore, phosphoenzyme formation by utilization of ATP in the presence of Ca 2ϩ proceeds very efficiently in the presence of Br 2 -TITU, and much higher levels of ADP-sensitive phosphoenzyme are reached in the presence than in the absence of Br 2 -TITU. We also demonstrated by direct measurements that phosphoenzyme formation by utilization of P i in the absence of Ca 2ϩ (i.e. E2P) is not inhibited by Br 2 -TITU but is in fact favored when the P i concentration is limiting. This latter effect may be due to looser A and P domain interaction (see below) and greater accessibility of the phosphorylation site to P i . Hydrolytic cleavage of E2P is only slightly inhibited by Br 2 -TITU.
The inhibitory effect of Br 2 -TITU is quite different from that of TG and DBHQ, inasmuch as these inhibitors interfere with Ca 2ϩ binding and enzyme phosphorylation with ATP or P i . It is known that TG produces a true dead-end complex, stabilizing a conformation equal or quite similar to E2 (10). Thereby, conversion to the E1-2Ca 2ϩ conformation in the forward direction of the cycle or to the E2P conformation in the reverse direction of the cycle is precluded by TG binding. On the contrary, Br 2 -TITU does not produce any stabilization of E2 because we found no hindrance to conversion either to E1Ca 2 when Ca 2ϩ was added or to E2P when P i was added in the absence of Ca 2ϩ . However, under conditions of ATP utilization in the presence of Ca 2ϩ , the inhibitory effect of Br 2 -TITU was realized by interference with the E1P-2Ca 2ϩ to E2P-2Ca 2ϩ transition, resulting in accumulation of E1P-2Ca 2ϩ . Thereby, formation of E2P was reduced. These direct measurements contradict previous inferences based on fluorescence measurements (1), which suggested that Br 2 -TITU inhibition results in accumulation of E2P by inhibiting its hydrolytic cleavage and is not likely related to interference with the E1P-2Ca 2ϩ to E2P-2Ca 2ϩ transition.
The experiments on proteinase K and trypsin digestion are indicative of A domain rotation and gathering of the ATPase cytosolic domains following removal of Ca 2ϩ to yield E2 (15,16). Further inclination of the A domain and compact gathering of the headpiece were observed upon addition of P i to yield E2P. We show here that Br 2 -TITU interferes significantly with positioning of the A domain and protection of the proteinase K digestion sites as expected in the E2 and E2P states of the ATPase. It is then apparent that in the presence of Br 2 -TITU, considerable fluctuation of the loops connecting the A domain to the transmembrane region is allowed even when the enzyme is placed in the E2 or E2P states by equilibration with EGTA or P i . Considering the involvement of A domain positioning and interactions with the P domain in specific phosphoenzyme states (6, 7), we suggest that the Br 2 -TITU conformational interference is related to the delay of the E1P-2Ca 2ϩ to E2P-2Ca 2ϩ transition observed in the kinetic experiments. This interference results in accumulation of E1P-2Ca 2ϩ and reduction of the steady state velocity of catalytic and transport activity. Therefore, the Br 2 -TITU inhibition mechanism is quite different from that of TG, which is based on strong stabilization of the compact headpiece conformation.